Data Flow Classification Device for Bufferbloat Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data flow classification methods, such as queue scheduling and active queue management, fail to effectively differentiate between high and low priority packets, leading to bufferbloat and adverse effects on throughput and user experience, especially for real-time applications like VoIP and video streaming.
Innovation Solution
A data flow classification device that uses a forwarding circuit and a configuring circuit to classify data flows based on traffic thresholds, adjusting the elephant-flow traffic threshold according to queue states, allowing for preferential discard of high-traffic 'elephant flows' to prevent bufferbloat by lowering their transmission priority.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If packets are classified according to user setting, then packet classification can be performed, but it is not convenient to users and hard to classify packets properly
Solution Approach 1:
The system automatically classifies packets by detecting elephant flows based on traffic patterns and queue states, eliminating the need for manual user configuration. The classification is performed self-service through automated threshold comparison and dynamic adjustment mechanisms
Solution Approach 2:
The system uses feedback from queue state monitoring and traffic information to dynamically adjust classification thresholds and improve classification accuracy over time, adapting to changing network conditions without user intervention
2Loss of time
If Active Queue Management (AQM) is used to reduce average queue length, then queuing delay can be reduced, but packets of high transmission priority and low transmission priority are discarded by the same probability, affecting throughput of high priority packets
Solution Approach 1:
The system applies different discard probabilities to different packet classes by identifying elephant flows and assigning them differentiated treatment. High priority packets maintain normal discard rates while low priority elephant flows experience higher discard rates, achieving local quality differentiation in queue management
Solution Approach 2:
The system segments packets into different categories (elephant flows vs. normal flows, high priority vs. low priority) and applies distinct queue management policies to each segment, enabling differentiated service quality for different packet types
3Reliability
If elephant flow classification is applied to prevent bufferbloat, then buffer space can be reserved for high priority packets, but the system complexity increases
Solution Approach 1:
The system changes the parameter of traffic threshold dynamically based on queue states to identify elephant flows. By adjusting the threshold parameter according to current network conditions, the system achieves adaptive classification without requiring complex algorithms
Solution Approach 2:
The system replaces complex manual classification mechanisms with automated detection based on traffic information and threshold comparison. The classification is achieved through systematic parameter comparison rather than complex mechanical or manual processes
Data Source
AI summary
Disclosed is a data flow classification device including a forwarding circuit and a configuring circuit. The forwarding circuit looks the classification of an input flow up in a lookup table according to the information of the input flow, tags the packets of the input flow with the classification, and outputs the packets to a buffer circuit. The configuring circuit receives and stores the identification and traffic information of multiple flows, and accordingly calculates the traffic of the multiple flows, wherein the multiple flows include the input flow. The configuring circuit further determines an elephant flow threshold according to a queue length of the buffer circuit and a target length, determines the classifications of the multiple flows according to the comparison between the traffic of the multiple flows and the elephant flow threshold, and stores these classifications in the lookup table.